Hannestad and Scherrer analysed self-interacting Warm and Hot Dark Matter and found power spectra in disagreement with previous papers. We argue that they include self-interactions in a weak-coupling approximation, as for photons after recombination. The ``tightly coupled'' approximation used for pre-recombination photons, and previous discussions of self-interacting HMD, would have been more appropriate. Their approximation generates a Boltzmann hierarchy leading to a stiff system of equations. Furthermore contamination by gauge modes could invalidate their results.